To calculate street closure times, we estimate that trains are an average of 7,000 feet long, and that an average train carries 13,750 tons of coal with 110 tons of coal in each of 125 rail cars. In the case of coal trains bound for Cherry Point only, we estimate a smaller number of long trains—at 8,500 feet, consistent with publicly available project documents—but this exception to our general rule has virtually no effect on aggregate crossing times. Sightline estimates existing coal shipments based on the most recent quarterly data provided by US Customs, though these figures may understate the actual number of coal trains.

We estimate that oil trains are the same length on average, 7,000 feet, carrying 70,000 barrels of crude oil with 700 barrels in each of 100 rail cars. Sightline makes detailed project-specific estimates of oil train volumes in “The Northwest’s Pipeline on Rails,” a report updated October 2013.

Our estimates of street closure times do not factor in additional time for sounding bells or closing and opening crossing gates, nor do we count street closure times that result from rail system capacity constraints, stalled trains, collisions, or derailments. Furthermore, we do not count coal trains bound for the Port of Morrow, Oregon nor oil trains bound for Port Westward, Oregon because it is possible that these trains may travel other routes or travel on Union Pacific rail lines. We also do not count oil trains bound for Longview where a speculative oil-by-rail project has been discussed.

We obtained data on rail grade crossings from queries of the US Federal Railway Administration’s (FRA) database files on rail grade crossings and accidents published by the Office of Safety Analysis. Query results were obtained from files current as of November 30, 2012 and current as of at least the end of 2012 for crossing inventory files. (One query we used frequently was for detailed information on public grade crossings, by state and county, and displayed by City; with separate entries for “In” a given City or “Near” a given City.) We describe only public grade crossings. The FRA data source is here.

All images were copied from Google Earth, then loaded into Microsoft Paint, where BNSF public rail grade crossings were designated with red rectangles. The view is toward the North in each image, unless otherwise noted. Descriptions of locations and structures were taken from Google Earth.

Addendum #1: One Northwest rail expert wrote to us to explain why our estimates are too conservative. Actual street closures are likely to be longer because:

…a big factor in the many towns and small cities with single track mains that have nearby sidings is the fact that two gate down crossings are always back to back. It is not random – because the natural operation of the rail system on single track mains is dependent on trains “crossing” (also called “meeting point”) at these points. The fact that one of the trains must pull out of the siding from a dead stop also means that its crossing times will be much longer…

What this means “in general” is that if a loaded coal train is waiting stopped in a siding then estimating 4:00 crossing times should be upgraded to 6:00 if the siding is close to the crossing. If it is further away (say 2 ½ miles like south of Cheney) then trains will be up to full speed (crossing times unaffected) but they will still cross back at Cheney closer in time (whatever the travel time to cross the 2 ½ miles is for the two trains).

  • Our work is made possible by the generosity of people like you!

    Thanks to Erik Magnuson for supporting a sustainable Cascadia.

  • I know this sounds like a nuance, but with high traffic rates it will put additional pressure on affected crossings and on solving the problem with expensive grade separated crossings. There are a lot of towns and small cities all the way back to PRB that are single track mains with sidings in or near them…

    This is because the capacity of a single track main is totally dependent on sidings being used effectively. A siding in the (time based) middle of a segment doubles the capacity of that segment. The additional capacity only comes if the train in the siding leaves immediately after the other train clears the switch…


    Other factors include radius (curvature) of track limits speed, local speed limits may be in effect, 150 car unit trains are 20% longer than current 125 car trains, location of existing and new sidings will affect crossing times and distribution (multiple crossings very close in time preventing backed up traffic from clearing).

    Addendum #2: Our street closure estimates assume that empty trains return by the same route that the full trains travel to reach their destinations. In the time since we conducted our analysis, new information has come to light suggesting that empties may return by alternate routes such as Stampede Pass.

    A technical appendix to the 2013 draft update of the Washington Transportation Department’s Rail Plan, with analysis conducted by Cambridge Systematics, says:

    Stampede Pass Route. This route underwent an operational change in mid-year of 2012, switching from bidirectional to paired operation with the Columbia River route between Pasco and the Puget Sound region. Under the new scheme, loaded westbound bulk trains take advantage of the lower grades along the river, and once emptied, return east over the Stampede Pass route.

    See page A-7, technical note 4a for more detail.

    If this analysis is correct and some of the empty Longview coal trains return by way of Stampede Pass then our estimates of street closure times should be lower for communities south of Longview, but higher for communities to the north.

    Addendum #3: In this series, we tried to choose reasonable train speeds for calculating street closure times. Here are a few of the reference points we relied on.

    • In 2012, the traffic consulting firm Parametrix evaluated trains in Seattle using direct observation. They found that, on average, trains at the crossings were traveling only 7.4 mph (at S. Holgate Street) and 8.1 mph (at S. Lander Street). The fastest they clocked any train during their study was less than 25 mph. For the purposes of their calculations, they assumed coal train speeds through Seattle of 20 mph. US DOT crossing inventories indicate that the maximum train speed at crossings in Seattle is 30 mph through North Waterfront district and 50 mph through SODO district.
    • In June 2012, this Oregonian article evaluating potential traffic congestion from coal trains in rural northwest Oregon estimated trains speeds of 25 mph at uncongested crossings, and 10 mph in congested areas.
    • In January 2013, this article at The Stranger asserted that coal trains in Washington would average 35 mph.